Authors: Sergei V. Chekanov, Wasikul Islam, Nicholas Luongo

Journal of High Energy Physics (In Review) • 2025 (Est.)

This paper explores different strategies for enhancing sensitivity to new heavy resonances that decay into two or more Higgs bosons. This is achieved using two neural network architectures: an unsupervised autoencoder for anomaly detection and a supervised classifier. The autoencoder is trained on a small fraction of Standard Model (SM) Monte Carlo simulated events to calculate the loss distribution for input events, aiding in determining the extent to which events can be considered anomalous. The supervised classifier uses the same inputs but is trained on events simulated using both beyond Standard Model (BSM) and SM processes. By applying selection cuts to the output scores, we compare the sensitivities of the two approaches.

Authors: Jackson Barr, Diptaparna Biswas, Maxence Draguet, Philipp Gadow, Emil Haines, Osama Karkout, Dmitrii Kobylianskii, Wei Sheng Lai, Matthew Leigh, Nicholas Luongo, Ivan Oleksiyuk, Nikita Pond, Sébastien Rettie, Andrius Vaitkus, Samuel Van Stroud, Johannes Wagner

Journal of Open Source Software • August 2025

High energy physics studies the fundamental particles and forces that constitute the universe, often through experiments conducted in large particle accelerators such as the Large Hadron Collider (LHC) (Evans & Bryant, 2008). Salt is a Python application developed for the high energy physics community that streamlines the training and deployment of advanced machine learning (ML) models, making them more accessible and promoting shared best practices. Salt features a generic multimodal, multitask model skeleton which, coupled with a strong emphasis on modularity, configurability, and ease of use, can be used to tackle a wide variety of high energy physics ML applications.

Authors: Sergei V. Chekanov, Wasikul Islam, Rui Zhang, Nicholas Luongo

Information • March 2025

A web-based tool called ADFilter (short for Anomaly Detection Filter) was developed to process collision events using autoencoders based on a deep unsupervised neural network. The autoencoders are trained on a small fraction of either collision data or Standard Model (SM) Monte Carlo simulations. The tool calculates loss distributions for input events, helping to determine the degree to which the events can be considered anomalous with respect to the SM events used for training. Therefore, it can be used for new physics searches in collider experiments. Real-life examples are provided to demonstrate how the tool can be used to reinterpret existing results from the Large Hadron Collider (LHC), with the goal of significantly improving exclusion limits. This tool is expected to mitigate the “reproducibility crisis” associated with various machine learning techniques, as it can incorporate machine learning approaches from third-party publications, making them accessible to the general public.

Authors: The ATLAS Collaboration

Journal of High Energy Physics • July 2023

A search for Higgs boson pair production in events with two b-jets and two τ-leptons is presented, using a proton–proton collision dataset with an integrated luminosity of 139 fb−1 collected at = 13 TeV by the ATLAS experiment at the LHC. Higgs boson pairs produced non-resonantly or in the decay of a narrow scalar resonance in the mass range from 251 to 1600 GeV are targeted. Events in which at least one τ-lepton decays hadronically are considered, and multivariate discriminants are used to reject the backgrounds. No significant excess of events above the expected background is observed in the non-resonant search. The largest excess in the resonant search is observed at a resonance mass of 1 TeV, with a local (global) significance of 3.1σ (2.0σ). Observed (expected) 95% confidence-level upper limits are set on the non-resonant Higgs boson pair-production cross-section at 4.7 (3.9) times the Standard Model prediction, assuming Standard Model kinematics, and on the resonant Higgs boson pair-production cross-section at between 21 and 900 fb (12 and 840 fb), depending on the mass of the narrow scalar resonance.

Authors: The ATLAS Collaboration

ATLAS Public Note • 2020

Separating charged and neutral pions as well as calibrating the pion energy response is a core component of reconstruction in the ATLAS calorimeter. This note presents an investigation of deep learning techniques for these tasks, representing the signal in the ATLAS calorimeter layers as pixelated images. Deep learning approaches outperform the classification applied in the baseline local hadronic calibration and are able to improve the energy resolution for a wide range in particle momenta, especially for low energy pions. This work demonstrates the potential of deep-learning-based low-level hadronic calibrations to significantly improve the quality of particle reconstruction in the ATLAS calorimeter.

Authors: Nicholas Luongo on behalf of the ATLAS Collaboration

15th Conference on the Intersections of Particle and Nuclear Physics - Madison, WI • June 2025

The ATLAS experiment at the Large Hadron Collider relies on advanced machine learning (ML) techniques to enhance the reconstruction and identification of physics objects, optimize detector performance, and improve data analysis strategies. In recent years, ML methods have been extensively developed and integrated across various domains. This talk presents an overview of recent advancements for Physics in ML applications within ATLAS.

Authors: Nicholas Luongo

APS Global Physics Summit 2025 - Anaheim, CA • March 2025

The ATLAS experiment at the Large Hadron Collider relies on advanced machine learning (ML) techniques to enhance the reconstruction and identification of physics objects, optimize detector performance, and improve data analysis strategies. In recent years, ML methods have been extensively developed and integrated across various domains. This talk presents an overview of recent advancements for Physics in ML applications within ATLAS.